INVALID BALLOTS IN DUHOK GOVERNORATE, IRAQ'S 2025 PARLIAMENTARY ELECTION: A DESCRIPTIVE ANALYSIS THROUGH BOUNDED RATIONALITY

Abstract

Invalid ballots are votes that cost the voter time and effort but are not counted, yet ballot-level evidence from Iraq and the Kurdistan Region is scarce. In Duhok Governorate, 5.55% of the ballots cast in the 2025 parliamentary election were invalid, slightly below the reported national share of about 6%. This study asks how these ballots are distributed across marking patterns and registration centers, and what a bounded rationality perspective can and cannot say about them. All 34,694 invalid ballots were classified from scanned images into seven patterns following the criteria of the Independent High Electoral Commission. The classification was carried out under the supervision of the author with five trainers of the same electoral office, and results are summarized with frequencies, percentages, and the minimum, median, and maximum share of each pattern across centers; no inferential tests are applied. Overvotes dominate: ballots marking three or more lists account for 42.00% of invalid ballots and those marking two lists for 39.38%, followed by blank ballots (7.96%), candidate number without a list (6.37%), marks outside the range (2.79%), marks between two lists (1.12%), and marks with a different pen (0.38%). Composition varies widely across the 41 registration centers, with the share of three-or-more-list ballots ranging from 0% to 72.76%, and an abrupt break between the first five centers and the rest cannot be separated from differences in classification. The patterns are heterogeneous, and the data cannot distinguish procedural error from protest. The results suggest that ballot format and voting instructions are a natural focus for further study. The findings are limited to one election in one governorate, contain no information on voter intent, and rely on internal coding.